Formation of Knowledge Bases of Fuzzy Regulators Based on the Physical Training Signal of a Nitrogen Cryogenic Plant Using a Genetic Algorithm

The developed earlier technology for the formation of knowledge bases of fuzzy regulators applied expert estimations, which led to underdeterminacy and subjectivity in control processes. The technology of knowledge extraction from a physically recorded learning signal using a genetic algorithm is described, which allows the formation of objective knowledge bases for fuzzy controllers of an intelligent control system. A comparison of different types of control models based on TANGO [1] is carried out. A method for optimal trajectories design as the gain coefficients schedule of the hybrid PID controller is presented. The effectiveness of the end-to-end information technologies application based on soft computing in intelligent control tasks is shown.